INDEX

Explanations

specific geographical locations and institution names

oai_token-act-pair · gpt-4o-mini Triggered by @bot

geographical regions and programming code

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 autorytatywna

-0.57

enares

-0.50

jsonwebtoken

-0.50

ptonshire

-0.49

niająca

-0.49

 saites

-0.49

ghis

-0.48

 foal

-0.48

bootstrapcdn

-0.48

 otomatig

-0.47

POSITIVE LOGITS

 Bretagne

0.72

Bretagne

0.69

 Cornwall

0.68

 Kerala

0.59

 Cornish

0.58

 Donegal

0.55

 Breton

0.53

xase

0.52

Kerala

0.51

 Tasmanian

0.51

Activations Density 0.066%

specific geographical locations and institution names

geographical regions and programming code

No Comments

No Known Activations